Nebula LLM

Nebula is a Large Language Model (LLM) for building generative AI applications from conversations.

πŸ“˜

Nebula is currently in private beta.

To request access, see the sign up form.

Nebula is a large language model (LLM) built by Symbl.ai. It is trained to perform generative tasks on human conversations. Nebula excels at modeling the nuanced details of a conversation and performing tasks on the conversation.

Nebula currently supports two model variations:

  • Chat - chatting with the model in a multi-turn fashion. Ideal when you're building an interface to chat with a model/assistant. You can use transcript and instructions as part of an input message and continue chatting with the model.
  • Instruct - single instruction and conversation transcript. Ideal when you're building a non-chat interface that only needs one task in isolation. You can use the transcript along with the instructions and generate output.

You can interact with Nebula using the Model Playground or the API. The Playground allows you to test the model with conversations and tasks without writing code, while the Model API enables you to interact with the model programmatically.

Nebula can analyze conversation transcripts and generate responses based on the conversation along with the instruction or question in the prompt. You can provide instructions to request summaries, follow-up questions, draft emails, issues to review, qualify sales leads, identify and recommend resolutions to customer issues, or even recommend business opportunities. You can also ask specific questions about the conversation as part of the instruction.

For example, you can instruct the Model API to analyze a conversation with a customer and identify themes, sentiments, and issues to address follow-up responses and sales opportunities. You can also adjust the generation parameters of the model that control characteristics such as diversity, randomness, and repetition to change the model's behavior based on your use case.

Use cases

Here are some use cases that include examples of instructions and questions that you can use in your prompts. For examples of using the prompts in an API request, see the quickstart.

Conversation analysis

Category Prompts
Conversation analysis
  • What are the actionables from this conversation?
  • What could be the customer's pain points based on the conversation?
  • Generate a summary for this conversation in bullet point format.
  • What are the follow-up items from this conversation?
  • What potential solutions/products could be offered based on the conversation?
Coaching and performance
  • Based on this conversation, what feedback or coaching points would you give to the employee who handled it?
  • What best practices can be derived from this conversation for future customer interactions?
Risk assessment
  • Are there any potential risks or issues identified in this conversation that we should be aware of?
  • Based on this conversation, is the customer likely to churn? If so, what might be the reasons?

Customer service

Category Prompts
Customer satisfaction and retention
  • Based on the conversation, how satisfied does the customer seem with the products/services?
  • What improvements or changes might the customer want, as per the conversation?
Follow up
  • Based on the conversation, what would be the best next step or follow-up action with this customer?
  • Could you draft a follow-up email to the customer based on our conversation addressing their concerns?
Customer behavior and preference analysis
  • What can we infer about the customer's behavior and preferences from this conversation?
  • What are the customer's priorities or decision-making criteria based on the conversation?
Policy and process improvement
  • Based on the customer's feedback in this conversation, are there any policies or processes we should reconsider or improve?
  • Does this conversation indicate any gaps or inefficiencies in our current workflows or procedures?
Crisis management
  • Did the customer express any critical issues or complaints in this conversation that require immediate attention?
  • Based on the conversation, what would be the most effective way to resolve the customer's issues and regain their satisfaction?
Building customer relationships
  • What opportunities for deepening the relationship with the customer can be identified from this conversation?
  • How can we personalize our interactions with the customer based on the details from this conversation?

Sales and business opportunities

Category Prompts
Customer segmentation
  • Based on this conversation, which customer segments does this customer most likely belong to?
Sales opportunities
  • What sales opportunities can be identified from this conversation?
Business opportunities
  • Are there any potential business opportunities or partnerships mentioned in this conversation?
  • Did the conversation reveal any unmet needs that could potentially lead to the creation of new products/services?
Product/service improvement
  • Based on this conversation, what potential improvements can be made to the product/service?
  • Did the customer highlight any features or aspects of our product/service that they particularly like or dislike?